ORIGINAL ARTICLE

Responses of the coastal bacterial community toviral infection of the algae Phaeocystis globosa

Abdul R Sheik1,4, Corina PD Brussaard2,3, Gaute Lavik1, Phyllis Lam1,5, Niculina Musat1,6,Andreas Krupke1, Sten Littmann1, Marc Strous1 and Marcel MM Kuypers1

1Max Planck Institute for Marine Microbiology, Celciusstra�e 1, Bremen, Germany; 2Department of BiologicalOceanography, NIOZ – Royal Netherlands Institute for Sea Research, Texel, The Netherlands and 3AquaticMicrobiology, Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, Amsterdam,The Netherlands

The release of organic material upon algal cell lyses has a key role in structuring bacterialcommunities and affects the cycling of biolimiting elements in the marine environment. Here weshow that already before cell lysis the leakage or excretion of organic matter by infected yet intactalgal cells shaped North Sea bacterial community composition and enhanced bacterial substrateassimilation. Infected algal cultures of Phaeocystis globosa grown in coastal North Sea watercontained gamma- and alphaproteobacterial phylotypes that were distinct from those in the non-infected control cultures 5 h after infection. The gammaproteobacterial population at this timemainly consisted of Alteromonas sp. cells that were attached to the infected but still intact hostcells. Nano-scale secondary-ion mass spectrometry (nanoSIMS) showed B20% transfer of organicmatter derived from the infected 13C- and 15N-labelled P. globosa cells to Alteromonas sp. cells.Subsequent, viral lysis of P. globosa resulted in the formation of aggregates that were denselycolonised by bacteria. Aggregate dissolution was observed after 2 days, which we attribute tobacteriophage-induced lysis of the attached bacteria. Isotope mass spectrometry analysis showedthat 40% of the particulate 13C-organic carbon from the infected P. globosa culture was remineralizedto dissolved inorganic carbon after 7 days. These findings reveal a novel role of viruses in theleakage or excretion of algal biomass upon infection, which provides an additional ecological nichefor specific bacterial populations and potentially redirects carbon availability.The ISME Journal (2014) 8, 212–225; doi:10.1038/ismej.2013.135; published online 15 August 2013Subject Category: Microbial ecosystem impactsKeywords: Alteromonas and Roseobacter; carbon remineralisation; nanoSIMS; Phaeocystis globosa;pyrosequencing; marine viruses

Introduction

Marine viruses are the most abundant entities anddynamic components of the microbial loop (Suttle,2005). Studies conducted in the last years havemade it increasingly evident that viruses are

significant driving forces in algal (Haaber andMiddelboe, 2009) and bacterioplankton populationsdynamics (Middelboe et al., 2003). Moreover,through the ‘viral shunt’ (Wilhelm and Suttle,1999), the release of dissolved organic carbon andnutrients from the particulate organic pool isenhanced, leading to increased substrate availabilityfor microbially mediated processes. Thereby, virusescan influence biogeochemical cycling in the world’soceans (Brussaard et al., 2005b; Haaber andMiddelboe, 2009).

The prymnesiophyte, Phaeocystis globosa are adominant algae with the ability to generate high-biomass spring blooms (Brussaard et al., 1996).Viruses infecting P. globosa (PgVs) have recentlybeen brought into culture, and virus–host interac-tions (for example, latent period) have been wellinvestigated (Baudoux and Brussaard, 2005).Viral-mediated lysis can account for up to 66% ofthe total mortality of P. globosa single cells(Baudoux et al., 2006) and can even control algalbloom formation (Brussaard et al., 2005a). In a

Correspondence: AR Sheik, Eco-Systems Biology Research Group,Luxembourg Centre for Systems Biomedicine, University ofLuxembourg, Campus Belval, 7, avenue des Hauts-Fourneaux,Esch-sur-Alzette L-4362, Luxembourg.E-mail: abdul.sheik@uni.lu4Current address: Eco-Systems Biology Research Group, Luxem-bourg Centre for Systems Biomedicine (LCSB), University ofLuxembourg, Campus Belval, 7, avenue des Hauts-Fourneaux,L-4362 Esch-sur-Alzette, Luxembourg.5Current address: National Oceanography Centre Southampton,Ocean and Earth Science, University of Southampton, WaterfrontCampus, European Way, Southampton, SO14 3 ZH, UK.6Current address: Department of Isotope Biogeochemistry, HelmholtzCentre for Environmental Research – UFZ, Permoserstra�e 15,Leipzig, Germany.Received 20 February 2013; revised 3 June 2013; accepted 14 July2013; published online 15 August 2013

The ISME Journal (2014) 8, 212–225& 2014 International Society for Microbial Ecology All rights reserved 1751-7362/14

www.nature.com/ismej

mesocosm study, it was shown that viral-mediatedlysis of P. globosa blooms may lead to rapid changesin the microbial community structure and enhancedbacterial carbon utilisation (Brussaard et al., 2005b).Hence, P. globosa is an ideal species to study theimpact of viruses structuring bacterial communitiesand in turn the transfer of algal biomass towardsmicrobial communities affecting coastal biogeo-chemical element cycling.

Past field observations have shown that thebacterioplankton communities during algal bloomsin the coastal North Sea waters are mainly dominatedby Alphaproteobacteria, Gammaproteobacteria andBacteroidetes (Brussaard et al., 2005b; Teeling et al.,2012). The gammaproteobacterial Alteromonadaceae(referred to as Alteromonas cells henceforth) andalphaproteobacterial Rhodobacteriaceae (referred toas Roseobacter cells henceforth) can become veryabundant during algal blooms (Pernthaler et al.,2001) and exhibit distinctive temporal patterns mostlikely in relation to the changes in the organic mattercomposition during the course of the algal blooms(Eilers et al., 2000).

Viral-mediated algal lysis may induce aggregateformation due to the release of lysis products andcan be associated with dense bacterial abundances(Brussaard et al., 2005b). Although the majority ofmarine pelagic bacteria exists as free-living cells, asubstantial portion lives attached to algal surfacesand aggregates (Azam et al., 1983). Aggregate-associated bacteria are often characterized by highcellular abundance, growth rate and enzymaticactivity relative to their free-living counterparts(Riemann and Grossart, 2008). It is estimated thatabout 37% of the aggregate-associated bacteria maybe killed by viral lysis because of the density ofpotential host cells within aggregates (Proctor andFuhrman, 1991). Moreover, bacterial cell lysis couldmediate aggregate dissolution. Consequently, virusesmight alter the efficiency of the biological carbon pumpby retaining dissolved organic matter (for example,carbon) within the euphotic zone (Brussaard et al.,2008). However, how viral lysis shapes the bacterialcomposition and diversity and how it influences thebacterial uptake of virally released organic compoundsand thereby mediating oceanic biogeochemical cycles,remain poorly understood.

The objective of the current study is to investigatethe effects of algal viral infection and subsequentlysis on bacterial community structure, and carbonand nitrogen transfer from the algae to thebacteiroplankton. The utilisation of 13C- and15N- labelled axenic biomass of the model algaeP. globosa biomass by bacterial communities fromthe coastal North Sea (0.8 mm pre-filtered) duringand after algal viral infection was followed by acombination of bulk stable isotopic and molecularanalyses as well as novel single-cell analyses.Catalysed reporter deposition – fluorescent in situhybridisation (CARD-FISH), together with ampliconpyrosequencing, was used to examine the changes

in the bacterial composition and diversity. Theapplication of high-resolution single-cell techniquesenabled us to visualise the occurrence of aggregatesand aggregate-associated and/or free-living cellsbacteria (atomic force microscopy imaging, AFM).Single-cell bacterial substrate assimilation wasfurther quantified using nanometer-scale secondary-ion mass spectrometry (nanoSIMS), whereas theorganic carbon remineralisation associated withbacteriophage lysis was quantified using isotope ratiomass spectrometry (IRMS).

Materials and methods

Generation of 13C- and 15N-labelled algal biomassAxenic cultures of P. globosa strain Pg G (A) wereobtained from the culture collection of RoyalNetherlands Institute for Sea Research (NIOZ). The13C- and 15N-labelled P. globosa biomass were generatedfrom exponentially growing axenic P. globosa culturegrown for 2 days in Erlenmeyer flasks in modifiedenriched artificial seawater (ESAW; Cottrell and Suttle,1991) containing 1mM H13CO3

� , 80mM of 15NO3� (as

sodium salts, 99 atom %, ISOTEC) and 5mM of PO43�

(as sodium salts, Sigma-Aldrich Chemie Gmbh,Munich, Germany). The cultures were grown under95mmol quanta m� 2 s�1 irradiance with a light to darkregime of 16:8h and at a temperature of 15±1 1C. Onday 3, cultures were centrifuged at 1500g (with a swingrotor) for 10 min to remove unincorporated 13C- and15N-labelled substrates from the media. Algal cellpellets formed after centrifugation were washed twiceand re-suspended in an ESAW media without nutrientloadings. The centrifuged and re-suspended algalcultures had similar sensitivity to viral infection whencompared with non-centrifuged control cultures, sug-gesting that physiology of P. globosa was unaltered bycentrifugation (data not shown).

P. globosa virus culturing and bacterial inoculumThe lytic P. globosa virus, PgV-07T (Baudoux andBrussaard, 2005) used in this study was producedfrom exponentially growing P. globosa cultures inESAW media. The bacterial populations used in thisexperiment were obtained from Southern North Sea,The Netherlands (December 2008). Before perform-ing experiment, North Sea water was filteredthrough 0.8-mm pore size filters (GTTP, 45 mm indiameter, Millipore, Eschborn, Germany) to excludeheterotrophic nanoflagellates and other zooplankton(no grazers were found during epifluorescencemicroscopy). About 25% of the total bacterial cellnumbers were excluded after 0.8 mm pre-filtration(data not shown).

Experimental set-upThe centrifuged and re-suspended 13C- and15N-labelled P. globosa cultures in ESAW mediawithout nutrient loadings were split into four

Single-cell bacterial uptake of viral lysatesAR Sheik et al

213

The ISME Journal

subcultures. Each of these subcultures were trans-ferred (10% v-v) to Erlenmeyer flasks containing 3 l1:1 mixture of f/2 (Guillard, 1975) and ESAW media.Bacterial inoculation, that is, 0.8-mm-pre-filtered NorthSea water (10% v/v) was then added to each of thesesubcultures. Two of these subcultures were theninfected with pre-filtered PgV-07T virus (0.2mm pore-size, Whatman) at an initial virus to algae ratio of 17:1.The other two cultures served as non-infected controlcultures and received medium instead of viral lysate inequal amount. Sampling for flow cytometric algal andviral abundance, bulk particulate 13C- and 15N-measurements, catalysed reporter deposition-fluores-cence in situ hybridization analyses (CARD-FISH) andnanoSIMS analyses were taken soon after addition ofviral lysate. Samples were taken frequently until day 1of the experiment (0, 1, 2, 3, 5, 8, 12, 19, 24 and 30 hpost addition of viral lysate), followed by dailysampling until day 7 of the experiment.

Control experimentsIndependent control experiments were performed inorder to investigate the influence of organics derivedfrom algal media and P. globosa viral lysates on thechanges in the North Sea bacterial abundance(Supplementary Figure 1). Experiments were con-ducted in four Erlenmeyer flasks (2 l) containing 1:1mixture of f/2 and ESAW media with North Seabacterial inoculation (10% v/v, 0.8mm pre-filtered).Two of these subcultures received pre-filteredPgV-07T virus (0.2mm pore-size, Whatman), whereasother two cultures received medium instead of virallysate in equal amount. Samples for flow cytometricbacterial abundance were taken frequently until day 1of the experiment (0, 3, 5, 8, 12, 19 and 24 h), followedby daily sampling until day 5 of the experiment.

AbundancesAlgal abundance was monitored using flow cytometrywith a Beckman Coulter EPICS XL-MCL benchtop flowcytometer, equipped with a 15 mW 488 nm argon laser.One millilitre samples taken at each time point werediluted up to 10-fold in autoclaved seawater (0.2mmfiltered). The flow cytometer trigger was set on the redchlorophyll autofluorescence (emission 4630 nm).

The abundance of P. globosa viruses and bacter-iophages were enumerated by using a 15 mW 488 nmargon laser Becton-Dickson FACSCalibur flowcytometer according to Brussaard (2004). In short,samples of 1 ml were fixed with 25% glutaraldehyde(0.5% final concentration, EM grade, Sigma- Aldrich,USA) for 15–30 min at 4 1C, flash frozen in liquidnitrogen and stored at � 80 1C until analysis.The thawed samples were diluted 50- to 1000-foldin sterile TE-buffer (pH 8.0) and stained withthe nucleic acid-specific dye SYBR Green I(Invitrogen-Molecular Probes, USA) at a finalconcentration of 0.5� 10�4 of the commercial stockfor 10 min at 80 1C. The flow cytometer trigger was

set on the green fluorescence, and data files wereanalysed as described by Brussaard et al. (2010).

CARD-FISH analysesCARD-FISH analyses were performed to identifyand quantify the bacterial populations following theprotocol of Pernthaler et al. (2004). Subsamplestaken at each time interval were fixed withparaformaldehyde (PFA, 1% final concentration)for 1 h at room temperature or overnight at 4 1C.Subsamples were filtered onto white polycarbonatemembrane filters (GTTP, 0.2 mm pore size,Millipore), washed with 5–10 ml of 1� phosphatebuffer saline (PBS), air-dried and stored at � 20 1Cuntil analysis. Samples were hybridised with thefollowing probes: Gamma42a for Gammaproteobac-teria together with Beta42a as a competitor (Manzet al., 1992), CF319a for Bacteroidetes (Manz et al.,1996), ALF986 for Alphaproteobacteria (Amann et al.,1997), Alt1413 for Alteromonas cells (Eilerset al., 2000), Ros593 for Roseobacter cells (Eilerset al., 2001). Hybridised filters were counterstainedwith 1mg ml� 1 of 4,6-diamidino-2-phenylindole(DAPI). Subsequently, all DAPI-stained and hybridisedcells were quantified using epifluorescence microscopy(Axioplan II Imaging, Zeiss, Jena, Germany). TheEUB-I-III probe set was also used as CARD-FISHcontrols for sample filters and yielded X82%DAPI-stained cells in all samples, but only thecounts for the more specific groups are presentedhere. More than 900 DAPI-stained cells and 450probe-specific hybridised bacterial cells were evalu-ated in about 25 randomly chosen microscopic fields.

Bulk carbon and nitrogen measurementsFor the determination of bulk particulate 13C- and15N- measurements, 30–80 ml of the experimentalcultures were filtered onto pre-combusted glass fibrefilters, (GF/F, Whatman) freeze dried and storedat room temperature until analysis. The C- andN-isotopic composition of particulate organicmatter was determined as CO2 and N2 released byflash combustion in an automated elementalanalyzer (Thermo Flash EA, 1112 Series; ThermoFisher Scientific, Schwerte, Germany) coupled to anisotope ratio mass spectrometer (Finnigan Deltaplus

XP, Thermo Scientific, Thermo Fisher Scientific).

Carbon remineralisationCarbon remineralisation was measured as dissolvedinorganic 13C-carbon (13C-DIC) from labelled bio-mass released within the plankton community inour incubation experiments. Subsamples (5 ml) werepoisoned with 30 ml of saturated mercury chloridesolution. The isotopic component of DIC wasdetermined after acidifying with 1% final concen-tration of hypo-phosphoric acid as described byAssayag et al. (2006) and was analysed on a gas

Single-cell bacterial uptake of viral lysatesAR Sheik et al

214

The ISME Journal

chromatography-isotope ratio monitoring massspectrometry (Optima Micromass, Manchester,UK). Any loss of 13C-DIC in the medium becauseof gaseous exchange with atmospheric CO2 wascorrected for as described in Sheik et al. (2013).

Ammonium analysisAmmonium was measured using a TrAAcs 800autoanalyzer (detection limit of 0.1 mM) on samples(B 5 ml) that were gently filtered through poly-sulfone filters (0.2-mm pore size, Acrodisc, GelmanSciences, Zwijndrecht, Netherlands) and stored at� 20 1C until analysis.

Atomic force microscopyAtomic force microscopy (AFM, NT-MDT Co.,Moscow, Russia) was performed to visualise surfacetopography of aggregates (Supplementary Figure 2)on formaldehyde-fixed samples that were filtered onto polycarbonate membrane filters (0.22 mm pore-size; Millipore). AFM analysis was performed in asemi-contact mode as described by Sheik et al.(2013) in air at scan rates between 0.5 and 1 Hz anda spring constant of 11.8 Nm�1. Images wereprocessed with flatten correction function of thesoftware (Nova P9, version 2.1.0.828; NT-MDT Co.,Moscow, Russia).

NanoSIMS analysesHalogen in-situ hybridization assay coupled tonanoSIMS (HISH-SIMS) (Musat et al., 2008) wasperformed to identify and quantify the substrateassimilation of individual Alteromonas (probeAlt1413) and Roseobacter (probe Ros593) cells.Enrichment of the 13C and 15N in the specificprobe-hybridised bacterial cells (Alt1413 andRos593, as indicated by their respective, high 19Fcontent derived from the fluorine containing OregonGreen-labelled tyramide used), was analysed withNanoSIMS 50 L (CAMECA, Paris, France). Theprimary ion beam had a nominal size between 50and 100 nm, and the sample was sputtered with adwelling time of 2 ms per pixel. The primary currentof the Csþ beam was 20–30 nA during acquisitionfor most images. For each analysis, we recordedsimultaneously secondary-ion images of naturallyabundant 12C (measured as 12C� ), 14N (measured as12C14N� ) and, similarly, 19F (halogen-labelled HISH-SIMS tyramide) for the identification of specificprobe-hybridised bacterial cells, and 13C and 15N forthe uptake quantification. NanoSIMS data sets wereanalysed using the Look@NanoSIMS software(Polerecky et al., 2012). Regions of interest (ROI)around individual bacterial cells were definedmanually on on the basis of 19F image. The isotoperatios (r¼ 13C/12C or 15N/14N) were calculated foreach ROI on the basis of the total 13C– and12C– counts for each pixel. Subsequently, the atomic

percentage of 13C and 15N were calculated as13C/(13Cþ 12C) and 15N/(15Nþ 14N), respectively.The 13C and 15N isotopic ratios served as a proxyto quantify the algal biomass that has been trans-ferred towards specific bacterial members.

Calculations of biovolume and single-cell assimilationof 13C and 15NBiovolume of Alteromonas and Roseobacter cellswere calculated in order to estimate the amount of13C and 15N algal biomass transfer due to viral lysisor non-infected algae. Epifluorescence microscopyimages taken during CARD-FISH analyses andbefore nanoSIMS analyses were used to determinethe dimensions of Alteromonas and Roseobactercells. Assuming these bacterial cells as rotationalellipsoids (Olenina et al., 2006), we deduced thebiovolume of Alteromonas and Roseobacter cells atthe 5 h (Alteromonas cells only), day 2 and day 7 ofthe experiment (Supplementary Table 1).

We quantified the 13C and 15N substrate assimilation(fmol per cell�1) within single cells of Alteromonasand Roseobacter due to P. globosa viral lysis, relativeto non-infected control cultures. This estimation wason the basis of the 13C and 15N enrichments ofAlteromonas and Roseobacter cells and calculatedbiovolume as described by Musat et al. (2008).

DNA extraction and amplicon pyrosequencingIn order to investigate the community response toviral infection, the same seawater was used in theinfected and non-infected cultures, and DNAsamples were taken at 6 h, day 2 and day 7 inparallel from both cultures. The DNA samples fromthe non-infected culture were used as a reference tomonitor the changes in microbial community due toviral infection in the infected culture. DNA extrac-tion was performed as described by Zhou et al.(1996) on samples that were filtered (100–200 ml)onto white polycarbonate membrane filters (GTTP,0.2-mm pore size, Millipore) and stored at � 20 1Cuntil analysis. The extracted DNA was furtherpurified using the Wizard DNA Clean-Up System(Promega Corporation, Madison, USA) as per themanufacturer’s instructions. The bacterial 16S rRNAgenes were amplified and sequenced using ampli-con pyrosequencing at the Research and TestingLaboratories (Lubbock, TX, USA). The pyrosequen-cing was performed at 6 h, day 2 and day 7 of theexperiment from the infected and non-infectedcontrol cultures targeting Gammaproteobacteria(forward primer 50-CMATGCCGCGTGTGTGAA-30,reverse primer 50-ACTCCCCAGGCGGTCDACTTA-30),Alphaproteobacteria (forward primer 50-ARCGAACGCTGGCGGCA-30, reverse primer 50-TACGAATTTYACCTCTACA-30) and Bacteroidetes (forwardprimer 50-AACGCTAGCTACAGGCTT-30, reverseprimer 50-CAATCGGAGTTCTTCGTG-30). The generatedsequences were processed and taxonomically

Single-cell bacterial uptake of viral lysatesAR Sheik et al

215

The ISME Journal

identified as per the company’s standard procedure(Sun et al., 2011) to the species level according tothe 497% sequence identity of 16S rRNA genes.Thereafter, the species percentage composition ofeach major bacterial group was based on the relativeabundance information within and among theindividual samples and relative numbers of reads(Supplementary Tables 2 and 3).

MEGAN 4, a metagenome analysing software, wasused to construct the heat plot of 16S rRNAamplicon sequence data set (SupplementaryFigure 3, (Huson et al., 2011)). Amplicon sequenceswere clustered with 497% sequence identity withUCLUST (Edgar, 2010), and BLASTN was used tocompare clustered sequences against the SILVArRNA database (http://www.arb-silva.de). The outputof this comparison was then parsed by MEGAN4(Huson et al., 2011) and mapped onto the NCBItaxonomy. The BLASTN comparison showed thatBacteroidetes 16S rRNA amplicon sequences targetedmostly uncultured Bacteroidetes species and hence isnot described further in the text. Comparison tool ofMEGAN4 software was used to generate a phylogenetictree from multiple data sets.

Statistical analysesUsing the Sigmastat version 3.5 software package(Systat Software GmbH, Erkrath, Germany), one-wayanalyses of variance (ANOVAs) were used to test fordifferences in the bacterial numbers (B25 micro-scopic fields of each replicate, n) and single-cell13C and 15N enrichments of Alteromonas and

Roseobacter cells in infected and non-infected P.globosa cultures at different time intervals. Further,the Pearson product moment correlation was used todetermine the correlation between algal cell num-bers, particulate 13C-carbon and 13C-DIC.

Results

Microbial abundancesAlthough the P. globosa cell abundance in non-infected control cultures increased by 10-fold in 4days (from 0.2 to 2.3� 106 cells ml� 1; Figure 1a),viral infection led to a decline of algal hostabundance from 18 h post infection (p.i.) onwards(from 0.2� 106 to 2.6� 103 cells ml� 1 by day 6).Correspondingly, the abundance of PgV in infectedcultures increased between 8 and 12 h p.i.(Figure 1b). The bacterial abundance in non-infectedP. globosa cultures increased steadily throughout the7-day experiment (2.2� 107 cells ml� 1 at day 7;Figure 1c). In infected cultures, bacterial abundanceincreased rapidly from the point of P. globosa celllysis (12 h) reaching a maximum by day 2 (2.1� 107

cells ml� 1), after which it dropped sharply (3.7� 106

cells ml� 1, by day 7, Figure 1c). The sharp declinein the bacterial numbers coincided with an increasein the number of bacteriophages (Figure 1d).

Changes in bacterial community compositionIn the non-infected control cultures, catalysed reporterdeposition– fluorescence in situ hybridization

Figure 1 Changes in microbial and viral abundances in response to viral infection of P. globosa. (a) algal abundance, (b) P. globosa virus(PgV-07T) abundance (c) total heterotrophic bacterial abundance and (d) bacteriophage. Closed symbols represent non-infected controland open symbols represent the virally infected cultures. Error bars indicate standard error of mean from duplicate batch cultures(s.e.m.).

Single-cell bacterial uptake of viral lysatesAR Sheik et al

216

The ISME Journal

(CARD-FISH) analysis revealed that Bacteroidetesand Alphaproteobacteria dominated by day 7,accounting for 56 and 34% of the total microbialpopulations, respectively (Figure 2a). Viral-mediated P. globosa lysis affected most strikinglythe gammaproteobacterial populations (Figure 2).Growth of Alteromonas cells was stimulated shortlyafter P. globosa viral infection, already at 5 and 8 hpost infection, yet before cell lysis of P. globosa(ANOVA, number of random microscopic fieldsanalysed (n)¼ 25, P¼o0.001). Independent controlexperiments to test for the effect of the viral lysateadded and potential organic metabolites excreted bythe algal host showed no significant increase inbacterial abundance until day 2 where a slightincrease was observed (Supplementary Figure 1).CARD-FISH and amplicon pyrosequencing analy-ses showed that Alteromonas dominated theGammaproteobacteria in the infected cultureswith a maximum abundance of 8.5� 106 cells ml�1

on day 2 (Figures 2c and 3). The Alphaproteobacteriaalso peaked at day 2 in the infected cultures (2.5� 106

cells ml�1). The majority of the Alphaproteobacteriaphylotypes in the infected cultures, as identified peramplicon pyrosequencing analysis, belonged todiverse community of Roseobacter-related organisms(Figure 3, Supplementary Table 2). In contrast,alphaproteobacterial phylotypes in the non-infectedcontrol cultures showed Leisingera sp. to bedominant (Figure 3, Supplementary Table 2). Likethe Alphaproteobacteria, the Bacteroidetes popula-tion increased towards the end of the experiment in

the non-infected algal cultures. In contrast,Bacteroidetes in the infected cultures were hardlyaffected, maintaining more or less stable butlow cell abundance throughout the experiment(Figure 2b). Epifluorescence microscopy revealedthat Alteromonas cells from the infected P. globosacultures showed (micro)aggregate association soonafter P. globosa cell lysis (Table 2). The presence of(micro)aggregates was confirmed by atomic forcemicroscopy (AFM) imaging (SupplementaryFigure 2). In contrast to Roseobacter, Alteromonascells showed an increasing degree of aggregationuntil day 2 of the experiment (Table 1). Roseobactercells were mainly found in aggregates towards theend of the experiment. Thereafter, gammaproteobac-terial populations and similarly Alteromonas cells ininfected P. globosa cultures dropped until day 5.Interestingly, amplicon pyrosequencing analysisindicated that only a single phylotype, Alteromonassp., dominated the gammaproteobacterial popula-tions after P. globosa viral lysis and persisted through-out the experiment (Figure 3, SupplementaryTable 3). However, in non-infected control cultures,phylotype Galciecola sp. dominated at 6 h and day 2.Thereafter, at day 7 gammaproteobacterial phylotypesbecame more diverse (Figure 3, SupplementaryTable 3).

Carbon remineralizationThere was no significant change in the particulateorganic 13C-carbon (13C -POC) and particulate

Figure 2 The abundance of major bacterial groups of Alphaproteobacteria, Gammaproteobacteria and the Bacteroidetes(a, b) as determined by CARD-FISH analyses due to viral lysis of P. globosa relative to the non-infected control cultures. Withinspecific taxonomic groups, P. globosa viral lysis led to rapid changes in the abundance of Alteromonas (Gammaproteobacteria)followed by Roseobacter (Alphaproteobacteria) cells (c, d). Please note the different y-axis scale in the upper and bottom panels.The inlay in the bottom panel (c, d) shows the changes in the abundance of Alteromonas and Roseobacter cells for the first 12 hof the experiment. An initial doubling in the abundance of Alteromonas cells at 5 h post infection is indicated by an arrow (d).Error bars indicate s.e.m.

Single-cell bacterial uptake of viral lysatesAR Sheik et al

217

The ISME Journal

organic 15N-nitrogen (15N -PON) content with time inthe non-infected control cultures (Figures 4a and b).However, viral lysis of P. globosa led to a markeddecline in the amount of 13C -POC, which correlatedstrongly with the declining cell abundances (thePearson product moment correlation, R¼ � 0.947,Pp0.001, Figure 4b). On the basis of net decline ofthe 13C-POC by day 7 of the experiment, around 44%of the P. globosa biomass from the infected cultureswas converted to dissolved organic forms (Figure 4a,Supplementary Table 4). Meanwhile, there was nosignificant decline in 15N-PON (ANOVA, P¼ 0.004,Figure 4a). The decrease in the 13C-POC in infectedP. globosa cultures strongly correlated with theamount of organic carbon remineralized to 13C-DIC(R¼ � 0.991, Pp0.001, Figure 4c). By day 1 of theexperiment, the amount of carbon remineralised inthe infected cultures was already substantiallyhigher compared with the control cultures (ANOVA,P¼o0.001). Comparing the net decrease in 13C-POCand the net increase in the 13C-DIC, 22.5 mM13C, equivalent to 40% of the shunted particulate13C-organic carbon was remineralized by day 7(Figure 4c, Supplementary Table 4). In contrast, inthe non-infected control cultures, 13C-DIC concen-trations did not significantly change until day 6when a small increase in 13C-DIC was observed.Throughout the experiment, there was no detectableammonium in both the non-infected and infectedP. globosa cultures (data not shown).

Effect of viral lysis on algal C and N transfer to specificbacterial groupsBy combining CARD-FISH and HISH-SIMS imaging,we measured the 13C and 15N enrichments of theAlteromonas and Roseobacter cells from the non-infected control and infected P. globosa cultures(Figures 5 and 6). In the non-infected P. globosacontrol cultures, the 13C and 15N enrichment andhence substrate assimilation of Alteromonas cellsremained low within the first 5 h (Figures 5a–d,6a and c and Table 2, Supplementary Table 4).In contrast, already by 5 h, Alteromonas cellsfrom the infected P. globosa cultures were character-ized by significant enrichment in 13C and 15N(Figures 5e–h, and 6a and c), with calculatedsubstrate assimilation rates of 2.14 fmol 13C cell�1

and 0.65 fmol 15N cell� 1 (P¼o0.001, Table 2).A maximum substrate assimilation of Alteromonasand Roseobacter cells from the non-infected controlcultures was observed at day 2 of the experiment. Atday 2 of the experiment, in the infected cultures,B18% of the bulk 13C-POC was assimilated by theAlteromonas biomass and was about 2.4-foldenhanced relative to Alteromonas biomass in thenon-infected cultures (Supplementary Table 4). The13C and 15N substrate assimilation of Alteromonasfrom the infected cultures P. globosa culturesdecreased from day 2 to day 7, whereas substrateassimilation of Roseobacter cells increased from day2 to day 7 (Table 2).

Table 1 Percentage of Alteromonas and Roseobacter cells as free-living and aggregate associated

Alteromonas cells

Non-infected cultures Infected cultures

Time n Free-living(mean±s.e.) %

n Aggregate-associated(mean±s.e.) %

n Free-living(mean±s.e.) %

n Aggregate-associated(mean±s.e.) %

0 h 15 100±0 0 0 24 100±0 0 05 h 21 100±0 0 0 35 78±2.5 26 22±3.918 h 19 99±0 0 1±0 324 49±3.6 335 51±2.6Day 2 67 81.3±2.7 15 18.6±2.4 168 12.2±2.1 1196 87.7±5.2Day 4 91 74.7±2.8 31 25.8±3.1 200 40.9±5.2 290 59.1±6.3Day 7 45 83.3±3.5 29 16.6±3.9 69 70.8±4.5 45 29.1±3.5

Roseobacter cells

Non-infected cultures Infected cultures

Time n Free living(mean±s.e.) %

n Aggregate associated(mean±s.e.) %

n Free living(mean±s.e.) %

n Aggregate associated(mean±s.e.) %

0 h 15 100±0 0 0 15 100±0 0 05 h 17 100±0 0 0 35 100±0 0 018 h 19 95.6±1.7 6 4.4±2.6 17 98.2±3.2 6 1.7±2.8Day 2 166 83.1±2.5 33 16.9±5.7 172 67.8±1.8 82 32.2±1.2Day 4 214 67.3±6.3 104 32.7±4.6 59 41.8±4.4 90 58.1±3.6Day 7 603 55.1±3.1 459 44.9±4.0 50 27.8±4.6 131 72.1±4.8

n¼ total number of cells per microscopic field.

Single-cell bacterial uptake of viral lysatesAR Sheik et al

218

The ISME Journal

Discussion

Role of algal viral lysis in structuring bacterialcommunity compositionThe distinct temporal patterns in the abundance ofmajor bacterial groups show that P. globosa virallysis led to significant and very rapid changesin the bacterial community composition. Further,viral-mediated P. globosa lysis promoted the growthof opportunistic Gamma- and Alphaproteobacteria(r-strategists) relative to slow growing bacteria (forexample, Bacteroidetes, K-strategists) (Fuchs et al.,2000). The bacterial community composition alsochanged in association with aging P. globosa cellsfrom non-infected control cultures. However, thesecommunities were distinct from those in the infectedcultures. The predominance of Alphaproteobacteriaand Bacteroidetes observed in the non-infectedcontrol cultures agrees well with observations fromfield studies performed during P. globosa blooms(Brussaard et al., 2005b; Alderkamp et al., 2006;Lamy et al., 2009).

Our results show that viral infections of P. globosaled to the rapid growth of Alteromonas cells, whichdominated the bacterial community soon afterP. globosa viral lysis. In contrast to the developmentof Alteromonas abundance, the relative contributionof Roseobacter cells rose slowly. The initial rapidgrowth of Alteromonas cells is consistent with

previous phytoplankton incubation studies (Allerset al., 2007, Sandaa et al., 2009). Although the rapidincrease in Alteromonas abundance coincided withhigher bacteriophage abundance, potential bacter-iophage lysis was apparently not able to prevent theinitial Alteromonas bloom.

Interestingly, Alteromonas populations in theinfected culture mainly consisted of a singlephylotype, Alteromonas sp., during the initialdoubling of Alteromonas abundance (6 h postinfection), and this phylotype remained a majorcomponent throughout the remainder of the experi-ment. Gammaproteobacterial phylotypes in thenon-infected control cultures were more diverse.The development of Methylophaga sp., a knowndimethylsulfide (DMS) degrader (Schafer, 2007), byday 7 in the infected cultures could be related to thepotential occurrence of P. globosa derived DMScompounds (Liss et al., 1994). Although we did notanalyse for DMS, its potential formation couldexplain the presence of this methylotrophicorganism in our cultures. Moreover, viral lysis ofP. globosa triggered the development of severalalphaproteobacterial phylotypes, most of whichbelonged to the Roseobacter group. In contrast toGammaproteobacteria, the alphaproteobacterialphylotypes in the non-infected control culturesconsisted mainly of a single phylotype, Leisingerasp. (affiliated to Roseobacter cells), which persisted

Figure 3 Temporal changes in the bacterial community composition of phylotypes belonging to Gamma- and Alphaproteobacteriaduring and after viral infection of P. globosa relative to non-infected control cultures.

Single-cell bacterial uptake of viral lysatesAR Sheik et al

219

The ISME Journal

and remained to be a major phylotype throughoutthe experiment. Overall, our results indicate that algalviral lysis structures bacterial community composi-tion by favouring distinct bacterial phylotypes.

In the coastal North Sea, the P. globosa blooms areoften associated with aggregates that are formed dueto the disintegration of P. globosa colonies and/orcell lysis (Brussaard et al., 2005b; Mari et al., 2005).Despite the presence of P. globosa single cells in thisstudy, both the atomic force and epifluorescencemicroscopy revealed that the initial increasingabundances of Alteromonas and to a lesser extentRoseobacter cells were aggregate associated. AFMimaging visualised that aggregate-associated bacteria

appeared to be retained within a flocculant-likematerial. Aggregate formation might be enhanceddue to the impediment in the release of P. globosastar-like structures upon viral infection, which hasbeen suggested to stimulate (micro)aggregate forma-tion (Sheik et al., 2013).

Although aggregation may enhance accessibilityto certain substrates and stimulate bacterial growth,the dense bacterial populations within aggregatesmay also enhance their chance contact with theirspecific phages, thereby stimulating successful viral(that is, bacteriophage) infection. The declining cellabundance of Alteromonas and Roseobactercells, which were mostly aggregate associated, inP. globosa-infected cultures, coincided with anincrease in ambient bacteriophage abundance.Assuming a burst size (number of viruses producedper cell) of 50 (Parada et al., 2006), the net decreasein the abundances of Alteromonas and Roseobactercells matched the net increase in the number ofambient bacteriophages.

Viral driven carbon and nitrogen flowIt is generally assumed that viruses stimulatebacterial growth by the release of organic substratesupon host cell lysis. Our nanoSIMS analyses reveala substantial carbon and nitrogen substrateassimilation by Alteromonas cells in the infectedcultures already by 5 h post infection, before theviral-induced cell lysis of P. globosa (latent period8–12 h, (Baudoux and Brussaard, 2005)). The stillintact P. globosa cells and the lack of extracellularP. globosa virus increase also show that viral lysis ofP. globosa had not occurred yet. Therefore, theisotopic enrichment in Alteromonas cells wouldhave to come from the leakage or enhanced excre-tion of organic compounds from the infected butstill intact P. globosa host cells. To our knowledge,leakage or enhanced excretion in response to viralinfection as an additional substrate source forbacterial growth, has not been reported previously.

Interestingly, at day 2, P. globosa viral lysisspecifically enhanced the single-cell 13C and15N assimilation of Alteromonas cells by about2.5-fold relative to Roseobacter cellular substrateassimilation. During this time, there was no majorchange in the bulk 15N-PON content (Figure 4). AsAlteromonas and Roseobacter cells are part of theparticulate organic matter, the 15N transfer fromalgae to bacteria does not influence the 15N-PONcontent. In the control cultures, small increases in13C and 15N content of both Alteromonas andRoseobacter cells at day 2 (Table 2) indicate thatthese bacteria at least partly utilised algal derived13C- and 15N-labelled substrates. The significantlylarger contribution of Alteromonas (B35%,Supplementary Table 4) and to a lesser extentRoseobacter cells (B6%) to particulate 13C-POCand 15N-PON in the infected cultures suggests anefficient transfer of P. globosa viral lysates towards

Figure 4 Changes in particulate organic (a) 15N-nitrogen (15N-PON),(b) 13C-carbon (13C-POC) and (c) carbon remineralization due toviral-mediated shunt. Closed symbols represent non-infectedcontrol and open symbols represent the viral-infected cultures.Error bars indicate s.e.m.

Single-cell bacterial uptake of viral lysatesAR Sheik et al

220

The ISME Journal

these specific bacterial members. The observeddifferences in substrate assimilation were mostlikely due to different substrate uptake affinities ofthese two genera. The Alteromonas cells are capableof utilising a high diversity of organic compoundsfor energy acquisition ranging from low molecularweight organics such as hexoses (Gomez-Consarnauet al., 2012) to complex substrates such as coralmucus (Allers et al., 2008). On the other hand,Roseobacter cells are known to prosper on the

phytoplankton-derived material such as algalosmolytes and monomers like amino acids(Zubkov et al., 2001). Thus, the cell abundancesattained by Alteromonas and Roseobacter and their13C and 15N substrate assimilation by day 2 indicatethat initial stages of P. globosa viral lysis favour thedevelopment of opportunistic bacteria, such asAlteromonas.

Bacterial utilisation might alter the virallyreleased P. globosa organic matter with time, from

Figure 5 NanoSIMS imaging visualising the transfer of algal biomass towards Alteromonas cells in non-infected control cultures (upperpanel) and infected P. globosa cultures (lower panel) at day 2 of the experiment. First column (a, e) illustrates the CARD-FISH imagetaken before nanoSIMS analyses. The corresponding CARD-FISH hybridised cells were located by the 19F signal during nanoSIMS (b, f)and their respective 13C (c, g) and 15N enrichments (d, h). Scale bar¼ 2 mm.

Figure 6 The 13C and 15N enrichments as deduced from nanoSIMS analyses indicate the transfer of isotopically labelled biomass toAlteromonas (a, c) and Roseobacter (b, d) cells in both non-infected control and infected P. globosa cultures. The 13C and 15N enrichmentsof Roseobacter cells at 5 h were not determined due to their low cell abundance.

Single-cell bacterial uptake of viral lysatesAR Sheik et al

221

The ISME Journal

readily available organic substrates to more refrac-tory compounds (Brussaard et al., 2005b). Given thehigh enzymatic activity of aggregate-associatedbacteria (Proctor and Fuhrman, 1991), we speculatethat the bacteriophage-mediated lysis might havefacilitated the enzymatic dissolution of aggregatesleading to enhanced organic carbon remineralisationrates. On the basis of our estimates, B40% of theparticulate 13C organic carbon was remineralized to13C-DIC by day 7 (Supplementary Table 4). The restlikely constituted of less labile particulate organiccarbon forms such as cellular P. globosa star-likestructures. In fact, the presence of Pseudoalteromonassp., an algal polysaccharide decomposer (Ivanovaet al., 2002) at day 7, might indicate the accumula-tion of more less labile particulate material.

In the infected P. globosa cultures, the 13C and15N single-cell substrate assimilation by Alteromonascells decreased by day 7, even though the particulate15N-PON remained unchanged. The reduced 13C and15N labelling of the Alteromonas cells at day 7 mightindicate that they utilised less labelled and lesslabile organic material like P. globosa star-likestructures (Sheik et al., 2013). In contrast, theincreased substrate assimilation of Roseobacter cellsby day 7 suggests that organic material released dueto potential Alteromonas phage lysis might have atleast partly favoured the observed enhanced growthof Roseobacter cells. The impact of such bacterioph-age-mediated lysis of the abundant bacterial specieson other bacteria in the marine environmentrequires further investigation that should includeidentification of specific bacteriophages.

In contrast, to the particulate 13C content, theparticulate 15N-nitrogen content did not show anysignificant change with time in the infectedP. globosa culture (Figure 4a). Assuming Redfieldstoichiometry (C/N¼ 6.6; (Redfield, 1934)), theobserved 25.5 mM increase in 13C-DIC would mean arelease of 3.6 mM of ammonium by day 7 due toremineralization of the isotopically labelled algae(Supplementary Table 4). However, no detectableregeneration of ammonium was observed. The lack

of detectable ammonium in our study could haveresulted from ammonium adsorption on aggregatesas suggested previously (Shanks and Trent, 1979).

Conclusions and implicationsOn the basis of our combined results, we proposethe following stages of P. globosa viral infections,where the availability of carbon and nitrogen frominfected host algae appeared to structure bacterialdiversity (Figure 7). During early hours of viralinfection, and still before cell lysis, P. globosa cellsleaked or excreted B20% of organic matter that didnot only stimulate substantial substrate assimilationby Alteromonas cells but also triggered its attach-ment to the infected algal host (stage 1). Thisassociation of bacteria with algae is known as thephycosphere (Bell and Mitchell, 1972). Phycosphereformation has been attributed to various environ-mental factors (Teeling et al., 2012). However, viralinfections stimulating algal leakage or excretion andpromoting the growth of the associated phycospherehas thus far not been reported.

The bacterial response to the algal viral infectionswas a rapid temporal succession of bacterial popula-tions consisting of distinct phylotypes. Moreover,our observations indicate that viral lysis ofP. globosa single cells resulted in the formation ofaggregates that were colonised densely with bacteria(stage 2). Differences in the size of aggregates,bacterial colonisation, time of occurrence andenvironmental factors (Mari et al., 2005) will deter-mine whether or not aggregate formation enhancesor impedes the biological pump. Subsequent aggre-gate dissolution due to potential bacteriophage lysisappeared to be responsible for the regeneration ofdissolved inorganic carbon (stage 3). The suddenappearance of r- strategists (such as Alteromonascells) and their rapid demise signifies the efficiencyof potential bacteriophage-mediated lysis and offersa potential explanation to their apparent rarity in theenvironment (Pedros-Alio, 2006). Cell lysis due tolytic viral infections is considered to be the simplest

Table 2 Single cell 13C and 15N substrate assimilation of Alteromonas and Roseobacter cells (f mol per cell) at various temporal stages ofP. globosa viral lysis relative to non-infected P. globosa cells

Alteromonas cells Roseobacter cells

Non-Infected cultures Infected cultures Non-Infected cultures Infected cultures

Time n f mol 13C cell �1

(mean±s.e.)f mol 15 N cell � 1

(mean±s.e.)n f mol 13C cell �1

(mean±s.e.)f mol 15 N cell �1

(mean±s.e.)n f mol 13C cell � 1

(mean±s.e.)f mol 15 N cell �1

(mean±s.e.)n f mol 13C cell � 1

(mean±s.e.)f mol 15 N cell � 1

(mean±s.e.)

0 h 8 0 0 12 0 0 ND — — ND — —5 h 12 0.03±0.01 0.03±0.02 21 2.14±0.16 0.05±0.04 ND — — ND — —Day 2 87 0.56±0.08 0.07±0.01 82 1.32±0.02 0.22±0.005 41 0.32±0.04 0.05±0.01 63 0.76±0.06 0.16±0.01Day 7 46 0.13±0.01 0.02±0.002 36 0.53±0.23 0.10±0.004 21 0.16±0.01 0.02±0.002 46 1.38±0.15 0.24±0.02

Abbreviation: ND, not determined.n¼number of single cells analysed by nanoSIMS.

Single-cell bacterial uptake of viral lysatesAR Sheik et al

222

The ISME Journal

explanation of how viruses structure bacterialcommunities and mediate global biogeochemicalcycles. Our results show that already before celllysis the leakage or excretion of organic matter byinfected yet intact algal cells shaped North Seabacterial community composition and enhancedbacterial substrate assimilation. If these resultscould be extrapolated to other algal species, loss oforganic matter by infected intact algae may have animportant but so far unrecognised role in globalcarbon and nitrogen cycles.

Conflict of Interest

The authors declare no conflict of interest.

Acknowledgements

We thank Gabriele Klockgether, Daniela Franzke, ThomasMax, Birgit Adam, Anna Noordeloos and Evaline vanWeerlee for their technical assistance. We are grateful totwo anonymous reviewers for their valuable suggestions,which helped us to improve this manuscript. We thankthe Max Planck Society (MPG) and Royal NetherlandsInstitute for Sea Research for financial support.

References

Alderkamp AC, Sintes E, Herndl GJ. (2006). Abundanceand activity of major groups of prokaryotic plankton inthe coastal North Sea during spring and summer.Aquat Microb Ecol 45: 237–246.

Allers E, Gomez-Consarnau L, Pinhassi J, Gasol JM,Simek K, Pernthaler J. (2007). Response of Alteromo-nadaceae and Rhodobacteriaceae to glucose andphosphorus manipulation in marine mesocosms.Environ Microbiol 9: 2417–2429.

Allers E, Niesner C, Wild C, Pernthaler J. (2008). Microbesenriched in seawater after addition of coral mucus.Appl Environ Microbiol 74: 3274–3278.

Amann R, Glockner FO, Neef A. (1997). Modern methodsin subsurface microbiology: in situ identification ofmicroorganisms with nucleic acid probes. FemsMicrobiol Rev 20: 191–200.

Assayag N, Rive K, Ader M, Jezequel D, Agrinier P. (2006).Improved method for isotopic and quantitativeanalysis of dissolved inorganic carbon in naturalwater samples. Rapid Commun Mass Spectrom 20:2243–2251.

Azam F, Fenchel T, Field JG, Gray JS, Meyer-Reil LA,Thingstad F. (1983). The ecological role of water-column microbes in the sea. Mar Ecol Prog Ser 10:257–263.

Baudoux AC, Brussaard CPD. (2005). Characterization ofdifferent viruses infecting the marine harmful algalbloom species Phaeocystis globosa. Virology 341:80–90.

Baudoux AC, Noordeloos AAM, Veldhuis MJW,Brussaard CPD. (2006). Virally induced mortality ofPhaeocystis globosa during two spring blooms intemperate coastal waters. Aquat Microb Ecol 44:207–217.

Bell W, Mitchell R. (1972). Chemotactic and growthresponses of marine bacteria to algal extracellularproducts. Biol Bull 143: 265–277.

Brussaard CPD, Gast GJ, vanDuyl FC, Riegman R. (1996).Impact of phytoplankton bloom magnitude on a pelagicmicrobial food web. Mar Ecol Prog Ser 144: 211–221.

Figure 7 Conceptual diagram illustrating the observed temporal microbial regulation and associated biogeochemical processes due toP. globosa viral lysis using Alteromonas CARD-FISH images as an example. Viral infection of P. globosa cells led to leakage or excretion ofalgal organic matter, stimulating substrate assimilation by Alteromonas cells and also triggered its attachment to the infected algal host(stage 1). Viral lysis of P. globosa single cells resulted in the formation of aggregates that were colonised mostly with Alteromonas cellswith an efficient transfer of viral lysates (stage 2). Potential bacteriophage lysis most likely led to aggregate dissolution, leading toregeneration of dissolved inorganic carbon and also less labile organic carbon (stage 3). Scale bar¼5 mm.

Single-cell bacterial uptake of viral lysatesAR Sheik et al

223

The ISME Journal

Brussaard CPD. (2004). Optimization of procedures forcounting viruses by flow cytometry. Appl EnvironMicrobiol 70: 1506–1513.

Brussaard CPD, Kuipers B, Veldhuis MJW. (2005a). Amesocosm study of Phaeocystis globosa populationdynamics: 1. Regulatory role of viruses in bloom.Harmful Algae 4: 859–874.

Brussaard CPD, Mari X, Van Bleijswijk JDL, Veldhuis MJW.(2005b). A mesocosm study of Phaeocystis globosa(Prymnesiophyceae) population dynamics—II.Significance for the microbial community. HarmfulAlgae 4: 875–893.

Brussaard CPD, Payet JP, Winter C, Weinbauer MG. (2010).Quantification of aquatic viruses by flowcytometry. In: SW Wilhelm, MG Weinbauer and CASuttle (eds) Manual of Aquatic Viral Ecology ASLOpp 102–109.

Brussaard CPD, Wilhelm SW, Thingstad F, Weinbauer MG,Bratbak G, Heldal M et al. (2008). Global-scaleprocesses with a nanoscale drive: the role of marineviruses. ISME J 2: 575–578.

Cottrell MT, Suttle CA. (1991). Wide-spread occurrenceand clonal variation in viruses which cause lysis of acosmopolitan, eukaryotic marine phytoplankter,Micromonas pusilla. Mar Ecol Prog Ser 78: 1–9.

Edgar RC. (2010). Search and clustering orders ofmagnitude faster than BLAST. Bioinformatics 26:2460–2461.

Eilers H, Pernthaler J, Glockner FO, Amann R. (2000).Culturability and in situ abundance of pelagic bacteriafrom the North Sea. Applied Environ Microbiol 66:3044.

Eilers H, Pernthaler J, Peplies J, Glockner FO, Gerdts G,Amann R. (2001). Isolation of novel pelagic bacteriafrom the German Bight and their seasonal contribu-tions to surface picoplankton. Appl Environ Microbiol67: 5134–5142.

Fuchs BM, Zubkov MV, Sahm K, Burkill PH, Amann R.(2000). Changes in community composition duringdilution cultures of marine bacterioplankton asassessed by flow cytometric and molecular biologicaltechniques. Environ Microbiol 2: 191–201.

Gomez-Consarnau L, Lindh MV, Gasol JM, Pinhassi J.(2012). Structuring of bacterioplankton communitiesby specific dissolved organic carbon compounds.Environ Microbiol 14: 2361–2378.

Guillard RRL. (1975). Culture of phytoplankton for feedingmarine invertebrates. Smith, Walter L, Chanley Matoira H(eds) Culture of Marine Invertebrate Animals Conference,Greenport, NY Oct, 1972 Viiiþ 338p. Illus PlenumPress: New York, NY, USA; London, England Isbn0-306-30804-5 29–60.

Haaber J, Middelboe M. (2009). Viral lysis of Phaeocystispouchetii: Implications for algal population dynamicsand heterotrophic C, N and P cycling. ISME J 3:430–441.

Huson DH, Mitra S, Ruscheweyh HJ, Weber N,Schuster SC. (2011). Integrative analysis of environ-mental sequences using MEGAN4. Genome Res 21:1552–1560.

Ivanova EP, Bakunina IY, Sawabe T, Hayashi K, AlexeevaYV, Zhukova NV et al. (2002). Two species ofculturable bacteria associated with degradation ofbrown algae Fucus evanescens. Microb Ecol 43:242–249.

Lamy D, Obernosterer I, Laghdass M, Artigas F, Breton E,Grattepanche JD et al. (2009). Temporal changes of

major bacterial groups and bacterial heterotrophicactivity during a Phaeocystis globosa bloom in theeastern English Channel. Aquat Microb Ecol 58: 95–107.

Liss PS, Malin G, Turner SM, Holligan PM. (1994).Dimethyl sulphide and Phaeocystis: a review. J MarineSyst 5: 41–53.

Manz W, Amann R, Ludwig W, Vancanneyt M,Schleifer K-H. (1996). Application of a suite of 16SrRNA-specific oligonucleotide probes designed toinvestigate bacteria of the phylum cytophaga-flavobacter-bacteroides in the natural environment.Microbiology 142: 1097–1106.

Manz W, Amann R, Ludwig W, Wagner M, Schleifer KH.(1992). Phylogenetic oligodeoxynucleotide probes forthe major subclasses of proteobacteria: problems andsolutions. Syst Appl Microbiol 15: 593–600.

Mari X, Rassoulzadegan F, Brussaard CPDD, Wassmann P.(2005). Dynamics of transparent exopolymericparticles (TEP) production by Phaeocystis globosaunder N- or P-limitation: a controlling factor of theretention/export balance. Harmful Algae 4: 895–914.

Middelboe M, Riemann L, Steward GF, Hansen V, Nybroe O.(2003). Virus-induced transfer of organic carbonbetween marine bacteria in a model community. AquatMicrob Ecol 33: 1–10.

Musat N, Halm H, Winterholler B, Hoppe P, Peduzzi S,Hillion F et al. (2008). A single-cell view on theecophysiology of anaerobic phototrophic bacteria.Proc Natl Acad Sci 105: 17861.

Olenina I, Hajdu S, Edler L, Andersson A. (2006).Biovolumes and size-classes of phytoplankton in theBaltic Sea. HELCOM Baltic Sea Environ Proc 106:1–144.

Parada V, oacute nica, Herndl GJ, Weinbauer MG. (2006).Viral burst size of heterotrophic prokaryotes in aquaticsystems. J Marine Biol Assoc UK 86: 613–621.

Pedros-Alio C. (2006). Marine microbial diversity: can itbe determined? Trends Microbiol 14: 257–263.

Pernthaler A, Pernthaler J, Amann R. (2004). Sensitivemulti-color fluorescence in situ hybridization for theidentification of environmental microorganisms. MolMicrob Ecol Manual 3: 711–726.

Pernthaler A, Pernthaler J, Eilers H, Amann R. (2001).Growth patterns of two marine isolates: Adaptations tosubstrate patchiness? Appl Environ Microbiol 67: 4077.

Polerecky L, Adam B, Milucka J, Musat N, Vagner T,Kuypers MMM. (2012). Look@NanoSIMS—a tool forthe analysis of nanoSIMS data in environmentalmicrobiology. Environ Microbiol 14: 1009–1023.

Proctor LM, Fuhrman JA. (1991). Roles of viral infection inorganic particle flux. Marine Ecol Progress SeriesOldendorf 69: 133–142.

Redfield AC. (1934). On the proportions of organicderivatives in sea water and their relation to thecomposition of plankton. James Johnstone MemorialVol 176: 92.

Riemann L, Grossart HP. (2008). Elevated lytic phageproduction as a consequence of particle colonizationby a marine Flavobacterium (Cellulophaga sp.).Microbial Ecol 56: 505–512.

Sandaa R-A, Gomez-Consarnau L, Pinhassi J, Riemann L,Malits A, Weinbauer MG et al. (2009). Viral control ofbacterial biodiversity – evidence from a nutrient-enriched marine mesocosm experiment. EnvironMicrobiol 11: 2585–2597.

Schafer H. (2007). Isolation of Methylophaga spp. frommarine dimethylsulfide-degrading enrichment

Single-cell bacterial uptake of viral lysatesAR Sheik et al

224

The ISME Journal

cultures and identification of polypeptides inducedduring growth on dimethylsulfide. Appl EnvironMicrobiol 73: 2580–2591.

Shanks AL, Trent JD. (1979). Marine snow: Microscalenutrient patches. Limnol Oceanograph 24: 850–854.

Sheik AR, Brussaard CPD, Lavik G, Foster RA, Musat N,Adam B et al. (2013). Viral infection of Phaeocystisglobosa impedes release of chitinous star-like struc-tures: quantification using single cell approaches.Environ Microbiol 15: 1441–1451.

Sun Y, Wolcott RD, Dowd SE. (2011). Tag-encoded FLXamplicon pyrosequencing for the elucidation ofmicrobial and functional gene diversity in anyenvironment. Methods Mol Biol (Clifton, Nj) 733:129–141.

Suttle CA. (2005). Viruses in the sea. Nature 437: 356–361.Teeling H, Fuchs BM, Becher D, Klockow C, Gardebrecht A,

Bennke CM et al. (2012). Substrate-controlled successionof marine bacterioplankton populations induced by aphytoplankton bloom. Science 336: 608–611.

Wilhelm SW, Suttle CA. (1999). Viruses and nutrientcycles in the sea. Bioscience 49: 781–788.

Zhou J, Bruns MA, Tiedje JM. (1996). DNA recovery fromsoils of diverse composition. Appl Environ Microbiol62: 316–322.

Zubkov MV, Fuchs BM, Archer SD, Kiene RP, Amann R,Burkill PH. (2001). Linking the composition ofbacterioplankton to rapid turnover of dissolveddimethylsulphoniopropionate in an algal bloom inthe North Sea. Environ Microbiol 3: 304–311.

Supplementary Information accompanies this paper on The ISME Journal website (http://www.nature.com/ismej)

Single-cell bacterial uptake of viral lysatesAR Sheik et al

225

The ISME Journal